Genmo

Founded by brothers Paras and Ajay Jain, both ex-Google Brain researchers, Genmo is building open-source foundation models for video generation. Its flagship release, Mochi 1, is a 10B-parameter Apache-licensed text-to-video model based on the company's novel Asymmetric Diffusion Transformer architecture, benchmarked competitively against Runway Gen-3 and Kling at launch. Genmo raised a $28.4M Series A led by NEA in 2024 with participation from The House Fund, Gold House Ventures, and WndrCo, positioning the company as the open-source counterweight in a closed-model-dominated category.

Genmo released the strongest openly-licensed video model to date, giving researchers and builders a credible alternative to closed APIs from Runway and OpenAI.

genmo.ai Foundation Models - Video & Image

United Statesprivate

At a glance

Total raised

$33.5M

last round series a

Headcount

11-50

approximate · public range

Headquarters

San Francisco

United States

Founded

2022

4 years active

Status

Private

Foundation Models - Video & Image

Latest valuation

$150M

Series A · 2024

For subscribers

Subscribe to see the full profile

Including Funding rounds, Bull / Bear thesis, Stock + earnings, Roster changes, Patents, News, and Open roles.

Founding

50% off

/month

94 of 100 spots left

Full

$10

/month

Available once all 100 Founding Member spots are claimed.

Get full access

Already subscribed? Sign in →

Funding history

Total raised

$33.5M

Rounds tracked

Latest stage

Series A

Latest valuation

$150M

Date	Stage	Amount	Valuation	Lead investors
Oct 22, 2024	Series A	$28.4M	—	NEA

Founder backgrounds

Ajay Jain

Co-founder & CTO

Earned a B.S. in computer science from MIT and a Ph.D. from UC Berkeley, where his research focused on diffusion models, neural radiance fields, and scalable generative systems. As a researcher at Google Brain and previously at Facebook AI Research, NVIDIA AI, and MIT CSAIL, he co-authored foundational papers including *Denoising Diffusion Probabilistic Models* (DDPM) and *DreamFusion* (text-to-3D from 2D diffusion), each cited tens of thousands of times. He co-founded Genmo in late 2022 to apply that lineage of generative modeling research to video. In late 2024, the lab released Mochi 1, the largest openly-available text-to-video model, reflecting the open-weight philosophy his prior work helped establish.

Paras Jain

Co-founder & CEO

Earned a PhD in EECS from UC Berkeley, advised by Ion Stoica and Joseph Gonzalez, and was affiliated with the RISE Lab, BAIR Lab, and Berkeley Deep Drive — the latter focused on autonomous vehicle research. He is a co-creator of the Denoising Diffusion Probabilistic Models (DDPM) paper and DreamFusion (text-to-3D using 2D diffusion), both foundational works in modern generative AI. Before founding Genmo, he turned down an offer from Tesla to remain in academia. He co-founded Genmo with his brother Ajay Jain. The company released Mochi 1, the largest openly-available text-to-video model, positioning itself as an open-source alternative in AI video generation.

AI bull / bear

AI bull / bear analysis hasn’t been generated for this company yet. Cases refresh quarterly using the latest classified articles and funding history.

In the news

Genmo · last 30 days

No articles ingested yet for Genmo. Once the hourly news pipeline is live, every article the classifier tags as mentioning this company appears here with its one-line AI summary and sentiment.

Leadership

Paras Jain

Co-founder & CEO

—

Ajay Jain

Co-founder & CTO

—

Products

Mochi 1
Genmo's flagship open-source text-to-video model, released October 2024 under an Apache 2.0 license. At 10 billion parameters, Mochi 1 is the largest openly-released video generation model and is built on Genmo's Asymmetric Diffusion Transformer (AsymmDiT) architecture, which decouples the text and visual streams to maximize visual capacity. The model produces 5-second 480p clips with strong motion quality and prompt adherence, and has become a popular base for community fine-tunes, research benchmarks, and commercial derivatives.
Genmo Studio
Genmo's hosted creative interface that lets users generate, iterate, and edit videos without running models locally. Studio exposes Mochi 1 and Genmo's earlier models through a web UI with prompt history, reference uploads, and side-by-side comparisons, plus higher-resolution and longer-duration variants only available in the cloud. The product serves as both a revenue channel and a feedback loop for improving the underlying open models, mirroring the Stability and Black Forest Labs playbook of pairing open weights with a hosted offering.
AsymmDiT Research Stack
Genmo's underlying research platform built around the Asymmetric Diffusion Transformer architecture, including training code, inference kernels, and evaluation harnesses released alongside Mochi 1 on GitHub. The stack is designed to be reproducible on commodity GPU clusters and has been adopted by academic groups and independent researchers building on top of Genmo's work. By releasing the full training recipe rather than only weights, Genmo positions itself as the open-research anchor of the video generation field.

Open roles

We don't have a live feed for this company's ATS. Their careers page has every open role.

View all careers ↗